The elderly account for an increasing number of injured patients seen in U.S. hospitals, and a disproportionally large share of the deaths. While geriatric injuries continue to increase in frequency, the ability to predict outcomes for this population has not improved. Specifically, we are unable to determine which patients will do well, which will benefit from specialized trauma center care, and which interventions are best suited to their specific health needs. This lack of knowledge extends to the system level, with lack of policy (such as pre-hospital triage protocols) to help guide care for this population. Current paradigms for studying injury were designed to predict outcomes in young, otherwise healthy patients. In a resilient young person, morbidity and mortality are driven by injury characteristics (injury mechanism, pattern, and severity) and quality of the care they receive. This model fails in elderly patients for two reasons: (1) outcomes in the elderly are function of many different variables; and (2) the model fails to take into account important contextual data about health and well-being in the months and years before an injury. Since injuries occur in the setting of pre-existing conditions and functional impairment, outcomes cannot be predicted by injury characteristics alone. To address this gap in knowledge, we will create a database that is needed to inform future intervention studies and trauma-related policies. Specifically, we propose the creation a nationally representative database of injured elderly patients to better define the relationship between health, function, and injury using the Health and Retirement Study (HRS) and HRS-linked Medicare data. Several reasons make this database compelling. First, it will allow us to combine injury outcome prediction models (based on ICD9-derived injury characteristics) with detailed health and functional status information. Second, due to its longitudinal nature, we will be able understand the contextual nature of health and function over time for the injured population. Third, because we will have Medicare claims data, we will have valuable detail regarding healthcare utilization after injury. Finally, it will allow us to conducta study in a cost-effective and timely manner that would be difficult to do prospectively. We propose three aims to better define the relationship between health, function, and injury. Aim 1 will describe the characteristics of injuries and functional status in the injured HRS-linked Medicare population. Aim 1 will also evaluate the association between worsening pre-injury function and the risk for sustaining an injury. Aim 2 will focus on post-injury outcomes. Specifically, Aim 2 will describe how injuries impact the trajectory of functional decline, and wil determine the relative contribution of patient and injury characteristics to poor outcomes. Finally Aim 3 will determine the patterns of healthcare utilization and risk factors for high costs in this population.